Other
PyTorch
android
qaihm-bot commited on
Commit
beedbd6
·
verified ·
1 Parent(s): 141f263

See https://github.com/qualcomm/ai-hub-models/releases/v0.61.0 for changelog.

Files changed (2) hide show
  1. README.md +55 -54
  2. release_assets.json +22 -22
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  library_name: pytorch
3
- license: other
4
  tags:
5
  - android
6
  pipeline_tag: other
@@ -14,7 +14,7 @@ pipeline_tag: other
14
  Geometry-guided Kernel Transformer is a machine learning model for generating a birds eye view represenation from the sensors(cameras) mounted on a vehicle.
15
 
16
  This is based on the implementation of GKT found [here](https://github.com/hustvl/GKT/).
17
- This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/gkt) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
18
 
19
  Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.
20
 
@@ -27,41 +27,41 @@ Below are pre-exported model assets ready for deployment.
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip)
31
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip)
32
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip)
33
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip)
34
- | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip)
35
- | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip)
36
- | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip)
37
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
38
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
39
- | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip)
40
- | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip)
41
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip)
42
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip)
43
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs8275.zip)
44
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip)
45
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa8775p.zip)
46
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs9075.zip)
47
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa7255p.zip)
48
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa8295p.zip)
49
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
50
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
51
 
52
  For more device-specific assets and performance metrics, visit **[GKT on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/gkt)**.
53
 
54
 
55
  ### Option 2: Export with Custom Configurations
56
 
57
- Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/gkt) Python library to compile and export the model with your own:
58
  - Custom weights (e.g., fine-tuned checkpoints)
59
  - Custom input shapes
60
  - Target device and runtime configurations
61
 
62
  This option is ideal if you need to customize the model beyond the default configuration provided here.
63
 
64
- See our repository for [GKT on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/gkt) for usage instructions.
65
 
66
  ## Model Details
67
 
@@ -76,35 +76,36 @@ See our repository for [GKT on GitHub](https://github.com/qualcomm/ai-hub-models
76
  ## Performance Summary
77
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
78
  |---|---|---|---|---|---|---
79
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 57.818 ms | 8 - 8 MB | NPU
80
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 103.807 ms | 7 - 7 MB | NPU
81
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 81.19 ms | 8 - 15 MB | NPU
82
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 207.877 ms | 8 - 22 MB | NPU
83
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 102.759 ms | 7 - 11 MB | NPU
84
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 109.148 ms | 7 - 13 MB | NPU
85
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 207.877 ms | 8 - 22 MB | NPU
86
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 110.873 ms | 7 - 10 MB | NPU
87
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 103.807 ms | 7 - 7 MB | NPU
88
- | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 71.522 ms | 1 - 12 MB | NPU
89
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 71.522 ms | 1 - 12 MB | NPU
90
- | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 58.937 ms | 1 - 8 MB | NPU
91
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 58.094 ms | 7 - 7 MB | NPU
92
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 104.019 ms | 7 - 7 MB | NPU
93
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 80.884 ms | 8 - 16 MB | NPU
94
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 199.987 ms | 7 - 17 MB | NPU
95
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 190.115 ms | 1 - 10 MB | NPU
96
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 109.889 ms | 8 - 10 MB | NPU
97
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 112.427 ms | 1 - 6 MB | NPU
98
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 112.427 ms | 1 - 6 MB | NPU
99
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 112.427 ms | 1 - 6 MB | NPU
100
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 199.987 ms | 7 - 17 MB | NPU
101
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 110.844 ms | 7 - 17 MB | NPU
102
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 104.019 ms | 7 - 7 MB | NPU
103
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 71.315 ms | 7 - 16 MB | NPU
104
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 190.115 ms | 1 - 10 MB | NPU
105
- | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 142.055 ms | 0 - 6 MB | NPU
106
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 71.315 ms | 7 - 16 MB | NPU
107
- | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 58.871 ms | 7 - 16 MB | NPU
 
108
 
109
  ## License
110
  * The license for the original implementation of GKT can be found
 
1
  ---
2
  library_name: pytorch
3
+ license: mit
4
  tags:
5
  - android
6
  pipeline_tag: other
 
14
  Geometry-guided Kernel Transformer is a machine learning model for generating a birds eye view represenation from the sensors(cameras) mounted on a vehicle.
15
 
16
  This is based on the implementation of GKT found [here](https://github.com/hustvl/GKT/).
17
+ This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/gkt) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
18
 
19
  Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.
20
 
 
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip)
31
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip)
32
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip)
33
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip)
34
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip)
35
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip)
36
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip)
37
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
38
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
39
+ | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip)
40
+ | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip)
41
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip)
42
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip)
43
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs8275.zip)
44
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip)
45
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa8775p.zip)
46
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs9075.zip)
47
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa7255p.zip)
48
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa8295p.zip)
49
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
50
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
51
 
52
  For more device-specific assets and performance metrics, visit **[GKT on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/gkt)**.
53
 
54
 
55
  ### Option 2: Export with Custom Configurations
56
 
57
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/gkt) Python library to compile and export the model with your own:
58
  - Custom weights (e.g., fine-tuned checkpoints)
59
  - Custom input shapes
60
  - Target device and runtime configurations
61
 
62
  This option is ideal if you need to customize the model beyond the default configuration provided here.
63
 
64
+ See our repository for [GKT on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/gkt) for usage instructions.
65
 
66
  ## Model Details
67
 
 
76
  ## Performance Summary
77
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
78
  |---|---|---|---|---|---|---
79
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 57.794 ms | 8 - 8 MB | NPU
80
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 103.379 ms | 7 - 7 MB | NPU
81
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 82.81 ms | 10 - 18 MB | NPU
82
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 210.778 ms | 8 - 21 MB | NPU
83
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 102.663 ms | 7 - 11 MB | NPU
84
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 113.996 ms | 8 - 13 MB | NPU
85
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 210.778 ms | 8 - 21 MB | NPU
86
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 112.086 ms | 7 - 10 MB | NPU
87
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 103.379 ms | 7 - 7 MB | NPU
88
+ | GKT | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 71.758 ms | 1 - 8 MB | NPU
89
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 71.758 ms | 1 - 8 MB | NPU
90
+ | GKT | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 59.217 ms | 2 - 9 MB | NPU
91
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 61.888 ms | 7 - 7 MB | NPU
92
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 104.653 ms | 7 - 7 MB | NPU
93
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 82.114 ms | 8 - 15 MB | NPU
94
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 199.12 ms | 8 - 20 MB | NPU
95
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 102.138 ms | 7 - 17 MB | NPU
96
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 190.133 ms | 0 - 10 MB | NPU
97
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 109.338 ms | 8 - 10 MB | NPU
98
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 112.636 ms | 0 - 10 MB | NPU
99
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 112.636 ms | 0 - 10 MB | NPU
100
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 112.636 ms | 0 - 10 MB | NPU
101
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 199.12 ms | 8 - 20 MB | NPU
102
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 110.854 ms | 7 - 17 MB | NPU
103
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 104.653 ms | 7 - 7 MB | NPU
104
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 69.951 ms | 7 - 16 MB | NPU
105
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 190.133 ms | 0 - 10 MB | NPU
106
+ | GKT | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 142.388 ms | 0 - 6 MB | NPU
107
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 69.951 ms | 7 - 16 MB | NPU
108
+ | GKT | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 58.601 ms | 7 - 16 MB | NPU
109
 
110
  ## License
111
  * The license for the original implementation of GKT can be found
release_assets.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "version": "0.60.0",
3
  "precisions": {
4
  "float": {
5
  "chipset_assets": {
@@ -9,13 +9,13 @@
9
  "qairt": "2.45.0.260326154327",
10
  "onnx_runtime": "1.27.1"
11
  },
12
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip"
13
  },
14
  "qnn_context_binary": {
15
  "tool_versions": {
16
  "qairt": "2.45.0.260326154327"
17
  },
18
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs8275.zip"
19
  }
20
  },
21
  "qualcomm-qcs8550-proxy": {
@@ -24,13 +24,13 @@
24
  "qairt": "2.45.0.260326154327",
25
  "onnx_runtime": "1.27.1"
26
  },
27
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip"
28
  },
29
  "qnn_context_binary": {
30
  "tool_versions": {
31
  "qairt": "2.45.0.260326154327"
32
  },
33
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip"
34
  }
35
  },
36
  "qualcomm-qcs9075": {
@@ -39,13 +39,13 @@
39
  "qairt": "2.45.0.260326154327",
40
  "onnx_runtime": "1.27.1"
41
  },
42
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip"
43
  },
44
  "qnn_context_binary": {
45
  "tool_versions": {
46
  "qairt": "2.45.0.260326154327"
47
  },
48
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_qcs9075.zip"
49
  }
50
  },
51
  "qualcomm-sa7255p": {
@@ -53,7 +53,7 @@
53
  "tool_versions": {
54
  "qairt": "2.45.0.260326154327"
55
  },
56
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa7255p.zip"
57
  }
58
  },
59
  "qualcomm-sa8295p": {
@@ -61,7 +61,7 @@
61
  "tool_versions": {
62
  "qairt": "2.45.0.260326154327"
63
  },
64
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa8295p.zip"
65
  }
66
  },
67
  "qualcomm-sa8775p": {
@@ -69,7 +69,7 @@
69
  "tool_versions": {
70
  "qairt": "2.45.0.260326154327"
71
  },
72
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_sa8775p.zip"
73
  }
74
  },
75
  "qualcomm-snapdragon-8-elite-for-galaxy": {
@@ -78,13 +78,13 @@
78
  "qairt": "2.45.0.260326154327",
79
  "onnx_runtime": "1.27.1"
80
  },
81
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
82
  },
83
  "qnn_context_binary": {
84
  "tool_versions": {
85
  "qairt": "2.45.0.260326154327"
86
  },
87
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
88
  }
89
  },
90
  "qualcomm-snapdragon-8-elite-gen5-for-galaxy": {
@@ -93,13 +93,13 @@
93
  "qairt": "2.45.0.260326154327",
94
  "onnx_runtime": "1.27.1"
95
  },
96
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
97
  },
98
  "qnn_context_binary": {
99
  "tool_versions": {
100
  "qairt": "2.45.0.260326154327"
101
  },
102
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
103
  }
104
  },
105
  "qualcomm-snapdragon-8gen1": {
@@ -108,13 +108,13 @@
108
  "qairt": "2.45.0.260326154327",
109
  "onnx_runtime": "1.27.1"
110
  },
111
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip"
112
  },
113
  "qnn_context_binary": {
114
  "tool_versions": {
115
  "qairt": "2.45.0.260326154327"
116
  },
117
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip"
118
  }
119
  },
120
  "qualcomm-snapdragon-8gen3": {
@@ -123,13 +123,13 @@
123
  "qairt": "2.45.0.260326154327",
124
  "onnx_runtime": "1.27.1"
125
  },
126
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip"
127
  },
128
  "qnn_context_binary": {
129
  "tool_versions": {
130
  "qairt": "2.45.0.260326154327"
131
  },
132
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip"
133
  }
134
  },
135
  "qualcomm-snapdragon-x-elite": {
@@ -138,13 +138,13 @@
138
  "qairt": "2.45.0.260326154327",
139
  "onnx_runtime": "1.27.1"
140
  },
141
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip"
142
  },
143
  "qnn_context_binary": {
144
  "tool_versions": {
145
  "qairt": "2.45.0.260326154327"
146
  },
147
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip"
148
  }
149
  },
150
  "qualcomm-snapdragon-x2-elite": {
@@ -153,13 +153,13 @@
153
  "qairt": "2.45.0.260326154327",
154
  "onnx_runtime": "1.27.1"
155
  },
156
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip"
157
  },
158
  "qnn_context_binary": {
159
  "tool_versions": {
160
  "qairt": "2.45.0.260326154327"
161
  },
162
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.60.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip"
163
  }
164
  }
165
  }
 
1
  {
2
+ "version": "0.61.0",
3
  "precisions": {
4
  "float": {
5
  "chipset_assets": {
 
9
  "qairt": "2.45.0.260326154327",
10
  "onnx_runtime": "1.27.1"
11
  },
12
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip"
13
  },
14
  "qnn_context_binary": {
15
  "tool_versions": {
16
  "qairt": "2.45.0.260326154327"
17
  },
18
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs8275.zip"
19
  }
20
  },
21
  "qualcomm-qcs8550-proxy": {
 
24
  "qairt": "2.45.0.260326154327",
25
  "onnx_runtime": "1.27.1"
26
  },
27
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip"
28
  },
29
  "qnn_context_binary": {
30
  "tool_versions": {
31
  "qairt": "2.45.0.260326154327"
32
  },
33
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip"
34
  }
35
  },
36
  "qualcomm-qcs9075": {
 
39
  "qairt": "2.45.0.260326154327",
40
  "onnx_runtime": "1.27.1"
41
  },
42
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip"
43
  },
44
  "qnn_context_binary": {
45
  "tool_versions": {
46
  "qairt": "2.45.0.260326154327"
47
  },
48
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_qcs9075.zip"
49
  }
50
  },
51
  "qualcomm-sa7255p": {
 
53
  "tool_versions": {
54
  "qairt": "2.45.0.260326154327"
55
  },
56
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa7255p.zip"
57
  }
58
  },
59
  "qualcomm-sa8295p": {
 
61
  "tool_versions": {
62
  "qairt": "2.45.0.260326154327"
63
  },
64
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa8295p.zip"
65
  }
66
  },
67
  "qualcomm-sa8775p": {
 
69
  "tool_versions": {
70
  "qairt": "2.45.0.260326154327"
71
  },
72
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_sa8775p.zip"
73
  }
74
  },
75
  "qualcomm-snapdragon-8-elite-for-galaxy": {
 
78
  "qairt": "2.45.0.260326154327",
79
  "onnx_runtime": "1.27.1"
80
  },
81
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
82
  },
83
  "qnn_context_binary": {
84
  "tool_versions": {
85
  "qairt": "2.45.0.260326154327"
86
  },
87
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
88
  }
89
  },
90
  "qualcomm-snapdragon-8-elite-gen5-for-galaxy": {
 
93
  "qairt": "2.45.0.260326154327",
94
  "onnx_runtime": "1.27.1"
95
  },
96
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
97
  },
98
  "qnn_context_binary": {
99
  "tool_versions": {
100
  "qairt": "2.45.0.260326154327"
101
  },
102
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
103
  }
104
  },
105
  "qualcomm-snapdragon-8gen1": {
 
108
  "qairt": "2.45.0.260326154327",
109
  "onnx_runtime": "1.27.1"
110
  },
111
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip"
112
  },
113
  "qnn_context_binary": {
114
  "tool_versions": {
115
  "qairt": "2.45.0.260326154327"
116
  },
117
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip"
118
  }
119
  },
120
  "qualcomm-snapdragon-8gen3": {
 
123
  "qairt": "2.45.0.260326154327",
124
  "onnx_runtime": "1.27.1"
125
  },
126
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip"
127
  },
128
  "qnn_context_binary": {
129
  "tool_versions": {
130
  "qairt": "2.45.0.260326154327"
131
  },
132
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip"
133
  }
134
  },
135
  "qualcomm-snapdragon-x-elite": {
 
138
  "qairt": "2.45.0.260326154327",
139
  "onnx_runtime": "1.27.1"
140
  },
141
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip"
142
  },
143
  "qnn_context_binary": {
144
  "tool_versions": {
145
  "qairt": "2.45.0.260326154327"
146
  },
147
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip"
148
  }
149
  },
150
  "qualcomm-snapdragon-x2-elite": {
 
153
  "qairt": "2.45.0.260326154327",
154
  "onnx_runtime": "1.27.1"
155
  },
156
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip"
157
  },
158
  "qnn_context_binary": {
159
  "tool_versions": {
160
  "qairt": "2.45.0.260326154327"
161
  },
162
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/gkt/releases/v0.61.0/gkt-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip"
163
  }
164
  }
165
  }